This application proposes five types of experiments. All are intended to develop improved behavioral methods for categorizing abused drugs, and/or for determining the abuse liability of drugs. All are based on the drug discrimination technique. Project 1 is applied in nature. It seeks to develop high specificity assay techniques which can determine whether newly developed antianxiety drugs produce subjective effects more like those of highly abused depressants such as ethanol, or more like those of less abused drugs such as diazepam. If successful, the method will provide a new preclinical method for predicting the abuse liability of anxiolytic drugs. Project 2 will investigate the degree to which behaviors learned in the no drug condition become contingent upon interoceptive cues present in that condition. Theory and experimental results presently disagree on this issue, which leads to incorrect interpretation of experimental results and hinders attempts to improve the drug discrimination procedure. Project 3 will test for the occurrence of stimulus masking in the drug discrimination preparation, to determine whether this is a frequent or rare phenomenon. Project 4 consists of three separate experiments. The first will calibrate the substitution test procedure by determining the minimum degree of cue overlap between the sensory effects of training and test drugs that can be detected by the procedure, and the percentage of drug-lever responses that occur with higher degrees of overlap. The next experiment will test for the occurrence of cue overlap between the sensory effects of drugs which produce pharmacologically dissimilar actions. The third experiment will determine the degree to which the specificity of the DD paradigm varies as a function of the training dosage employed. Project 5 will investigate the utility of several new multiple-drug training procedures which are designed to allow rapid measurements of the subjective effects of drugs. A total of 14 separate experiments are proposed using rats and pigeons as subjects.